A PAN architecture may have a hysteresis in network topology to prevent thrashing when there is a temporary disruption in the network, such as a temporary loss of backhaul connection. However, some systems and devices cannot tolerate a temporary loss of backhaul connection and need to switch PANs once the backhaul connection becomes unavailable.
Generally, a root node does not routinely communicate the status of the backhaul connection to the other nodes in the PAN. Instead, each node has to determine the state of the backhaul connection by sending upper layer messages and receiving a response from the backhaul. Since a node cannot send an upper layer message until it joins a PAN, a node cannot determine the status of the backhaul connection prior to joining. A node that requires a backhaul connection may join a PAN and then determine that a backhaul connection is unavailable.
A node that determines that the backhaul connection for its current PAN is unavailable, may determine that it needs to switch PANs. However, the node may not have direct visibility to another PAN, which may delay the switch to a different PAN.
When a node switches PANs, any child nodes remain with the current PAN by finding a new parent node or switch to the new PAN by unjoining the current PAN and joining the new PAN. This process is time consuming and causes the child node to be unavailable until it locates a new parent or completes the unjoining/joining process.